摘要
为了提高粒子群优化算法的收敛速度和搜索精度,提出一种基于粒子适应值来设置动态收敛因子的方法(FPSO),根据粒子当前适应值和目标值之间差值的变化设置动态收敛因子,弥补了当前仅靠经验设置收敛因子的不足,实现了适应值对收敛因子的有效扰动。同时,提出一种基于维度最大位移量的局部优化激活方法,使得新算法能快速的从局部优化中跳出。通过4个经典函数对改进的算法进行测试,结果表明了改进后算法的有效性。
In order to improve the convergence velocity and accuracy, the dynamic convergence factor based on the fitness value is introduced into the improved algorithm（FPSO） according to the difference between the present fitness value and the target value.The new algorithm corrects the defect that the setting of the dynamic convergence factor just depends on the experience only, and the dynamic convergence factor is disturbed effectively by the fitness value in the FPSO.Then, the method based on the max displacement on dimensions is presented, and it makes the new algorithm can jump out form the local optimization quickly.The validity of the improved algorithm is proved by four classical functions.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第6期1283-1285,1289,共4页
Computer Engineering and Design
关键词
粒子群优化算法
适应值
动态收敛因子
维度最大位移
局部优化激活
particle swarm optimization algorithm
fitness value
dynamic convergence factor
max displacement on dimensions
local optimization activation
